Epicardial adipose tissue (EAT) is the visceral fat located between the outer layer of the myocardium and the visceral pericardium [1
]. EAT volume (EATV) is correlated with various cardiovascular risk factors, independent of abdominal visceral adiposity, body mass index (BMI), hypertension, and diabetes mellitus [5
]. Two population-based studies, the Multi-Ethnic Study of Atherosclerosis and the Framingham Heart Study, showed that EATV is an independent risk predictor for cardiovascular disease [5
]. EAT is shown to be metabolically active and the source of pro-atherogenic mediators and adipocytokines [1
]. Because EAT and the myocardium are located close anatomically, it is predicted that cytokines/adipocytokines produced by infiltrated macrophages or by adipocytes could locally modulate myocardial function or contribute to the pathogenesis of coronary atherosclerosis [1
]. Recently, we [10
] and others [12
] showed that proinflammatory cytokines and adipocytokines are expressed and secreted at a higher level in the adipose tissue of individuals with coronary artery disease (CAD) than in individuals without CAD.
Abdominal fat distribution is dissimilar between men and women: Visceral fat obesity is the dominant form in men, while subcutaneous fat obesity is the dominant form in women [13
]. However, gender differences in EATV distribution and the influence of EAT on coronary atherosclerosis has never been considered. In this study, we evaluated gender disparities in EATV and its impact on coronary atherosclerosis by using 256-slice multi-detector computed tomography (MDCT).
Subjects and methods
We recruited 119 consecutive subjects who underwent 256-slice MDCT coronary angiography between October 2009 and April 2011 at the Kawashima Hospital, Tokushima, Japan. The subjects underwent MDCT if they had atherosclerotic risk factors such as age ≥65 years [15
], hypertension, smoking, diabetes mellitus, or dyslipidemia or symptoms suggestive of angina pectoris. According to the 2010 Appropriate Use Criteria for Cardiac Computed Tomography [16
] guidelines, cardiac CT is not necessarily recommended for asymptomatic individuals with low-to-moderate CAD risk. However, the prevalence of CAD was not negligible even in asymptomatic subgroups [17
]; hence, we employed MDCT in subjects with moderate-to-high CAD risk after they were informed of the radiation-exposure related risk, and they provided written informed consent. Coronary CT angiography was performed using a 256-slice scanner (Brilliance iCT, Philips Healthcare, Amsterdam, Netherlands), and the diagnostic accuracy of the coronary CT angiography for coronary luminal narrowing was validated by routine invasive coronary angiography. Exclusion criteria included a history of cardiac surgery, iodine-based contrast allergy, or renal failure (creatinine, >1.5 mg/dL). Hypertension was defined as a systolic blood pressure of ≥140 mm Hg and/or diastolic blood pressure of ≥90 mm Hg or as the current use of antihypertensive treatment. Diabetes was defined as fasting plasma glucose of ≥6.99 mmol/L (126 mg/dL) or the current use of hypoglycemic treatment. Hyperlipidemia was defined as fasting serum LDL-cholesterol of ≥3.62 mmol/L (140 mg/dL), HDL-cholesterol of <1.03 mmol/L (40 mg/dL), triglyceride of ≥1.58 mmol/L (140 mg/dL), and/or the current use of antihyperlipidemic treatment. The subjects were segregated into 2 groups: CAD group (presence of plaques resulting in >50% luminal narrowing in the major coronary arteries) and non-CAD group (no plaque or plaques resulting in ≤50% luminal narrowing). All subjects in the CAD and non-CAD groups were further segregated into younger (age, <65 years) and older subgroup (age, ≥65 years).
Multi-detector CT scan protocol
The Cardiac MDCT acquisition was performed with retrospective ECG-gated cardiac imaging. The MDCT scan was performed using the following parameters: detector collimation of 2 × 128 × 0.625 mm, creating 256 overlapping slices of 0.625-mm thickness via a dynamic z-flying focal spot, gantry rotation time of 0.27 s, and tube voltage 120 kVp. A current of 800–1050 mA (depending on patient habitus) was used for helical acquisitions and a current of 200 mA for axial acquisitions. Computed tomography dose index volume (CTDIV) was calculated as 74.1 ± 12.6 mGy (n = 50). The raw scan data were reconstructed with 75% of RR wave or particular optimal phase. A bolus dose of the contrast medium (Iohexol [Omnipaque; Daiichi-Sankyo Pharmaceutical, Tokyo, Japan], containing 350 mg iodine/mL) was injected at 0.7 mL/kg body weight within 9 s. Nitroglycerin (0.3 mg) was administered to all subjects immediately before CT imaging, and an oral β blocker (metoprolol, 60 or 120 mg) was administered 1 h before CT imaging to render heart rates <65 beats/min, if required.
Analysis of EATV
We performed volumetric quantification of EAT measured by the 256-slice MDCT, as described with some modifications [18
] (Figure ). Quantification of total EAT area (cm2
) was performed at a workstation (Real-INTAGE, Kubota, Japan) with dedicated software. Volumetric measurements were performed on axial views of 0.625-mm slice thickness and number of slices ranging between 300 and 320. The superior border of the EATV measurements was the lower surface of the left pulmonary artery origin, while the inferior border was the left ventricular apex. The EAT area around the proximal, middle, and distal segment of the major coronary arteries was included in the volumetric measurements. The EAT area was calculated by tracing a region of interest (ROI), which included the heart and EAT. The ROI was manually placed outside the line of the visceral pericardium on a cross-sectional axial image (yellow line, Figure ). The area outside the traced pericardium was excluded. A density range of -600 to -20 Hounsfield Units was used to isolate the adipose tissue. The EAT area of each slice was then summed and multiplied by the slice thickness and number of slices to determine the total EATV (cm3
Figure 1 Total epicardial adipose tissue volume (EATV) measurements on 256-slice MDCT. (A)–(C): Axial images. A region of interest (ROI) was manually placed along the visceral pericardium (yellow line) (A) and EAT was extracted on an axial image (green) (more ...)
Assessment of coronary atherosclerosis
The presence of coronary atherosclerosis was estimated by CT angiographical data using a 256-slice Brilliance iCT (Philips Medical Systems). Coronary arteries were assessed for luminal narrowing by using 3D visualization tools after the axial images were reviewed for determination of anatomy, quality of the study, and appearance of the vessels. The coronary artery tree was segmented according to the modified American Heart Association classification [23
]. Coronary vessel and diameter were assessed on 2D multiplanar reconstruction (MPR) and 2D thin slab maximum intensity projection (MIP) images. A 3D volume rendered (VR) image was used to display long segments of the vessels and their branches. To compare the possible gender differences in the calcified lesion, we compared the Agatston scores between the men and women in CAD and non-CAD [24
Values are expressed as mean ± SD unless otherwise indicated. For comparison of the mean in the 2 groups, we used unpaired t-test when samples were normally distributed and non-parametric Mann-Whitney's U test when samples were not normally distributed. χ2 test was used to examine differences with categorical variables. Multigroup comparisons of variables were performed using one-way ANOVA followed by Tukey-Kramer HSD (honestly significant difference) test. Multiple logistic regression analysis was performed to adjust confounding factors. Variables were treated as continuous: one with a risk as 1 and one with no risk, i.e., as 0. We investigated the independent variables for detecting coronary artery luminal narrowing (>50%) by using unadjusted (univariate) and adjusted style (multivariate) for age, BMI, and other established risk factors (hypertension, hyperlipidemia, and diabetes mellitus). All analyses were performed using Jump version 9.0.2 software (SAS Institute Inc., Cary, NC). P values less than 0.05 were considered significant.
Of the 119 subjects who underwent cardiac MDCT, 29 were excluded from analysis. Among the men, 11 were excluded because of differences in slice levels and 2 because of insufficient image quality; among the women, 11 were excluded because of differences in slice levels and 5 because of insufficient image quality. A total of 90 subjects (men: n = 47; age = 63 ± 12 years; women: n = 43, age = 64 ± 12 years) were analyzed (Table ). By using MDCT, 22 men and 16 women were segregated into the CAD group (>50% luminal narrowing), while 25 men and 27 women were segregated into the non-CAD group. Among the 47 men, 36 (77%) had normal weight (BMI < 25 kg/m2), 8 (17%) were overweight (BMI 25–30 kg/m2), and 3 (6%) were obese (BMI > 30 kg/m2). Among the 43 women, 33 (77%) had normal weight (BMI < 25 kg/m2), 7 (16%) were overweight (BMI 25–30 kg/m2), and 3 (7%) were obese (BMI > 30 kg/m2). When the non-CAD and CAD groups were combined, we found that the EATV was higher in men than in women (80 ± 33 vs. 65 ± 21 cm3; p = 0.0089), but the mean EATV/height and EATV/BSA were comparable (Table ).
Characteristics of the study population
Distribution of BMI and EATV/BSA in non-CAD and CAD subjects
EATV, EATV/height, and EATV/BSA were considerably higher among men in the CAD group than in those in the non-CAD group (Figure B-D) but not significantly different among women in the non-CAD and CAD groups (Figure B-D). BMI was not different in both groups (Figure A). Because age was higher in the CAD group than in the non-CAD group (Table ), we tried to minimize the confounding effects of age by comparing patients in the younger (<65 years) and older (≥65 years) subgroups. EATV, EATV/height, and EATV/BSA were higher among men in the CAD group than in the non-CAD group both in younger (<65 years) and older (≥65 years) subgroups; however, in women, no significant difference was found in these values (Additional file 1
and Additional file 2
Figure 2 Comparison of BMI (A), EATV (B), EATV/height (C), and EATV/BSA (D) in non-CAD (○) and CAD (●) subjects. BMI, body mass index; EATV, epicardial adipose tissue volume; BSA, body surface area. Coronary artery disease (CAD) was defined if (more ...)
Next, we examined the relationship between the degree of coronary artery stenosis and EATV. Subjects were divided into the following categories: grade 0 = no plaques in the major coronary branches; grade 1 = ≤25% luminal narrowing; grade 2 = ≤50% luminal narrowing; grade 3 = >50% luminal narrowing. EATV/BSA was larger among men in grade 2 and grade 3 than in grade 1 (Figure ). In men and women, the Agatston score did not differ between the non-CAD and CAD groups (Figure , upper panel). There was no correlation between EATV/BSA and Agatston score in the non-CAD and CAD groups (Figure , lower panel).
Figure 3 EATV/BSA in subjects with degrees of coronary luminal stenosis. Subjects were segregated into the following categories based on the degree of coronary luminal stenosis determined using MDCT: grade 0 = no plaque in the major coronary arteries; grade 1 (more ...)
Figure 4 Comparison of Agatston score in Non-CAD (○) and CAD (●) subjects and linear correlation between EATV/BSA and Agatston score in men and women. EATV, epicardial adipose tissue volume; BSA, body surface area. Coronary artery disease (CAD) (more ...)
Correlation between EATV/BSA and variables
When non-CAD and CAD groups were combined, EATV, EATV/height, and EATV/BSA were not correlated with BMI (data not shown). Even after segregation into non-CAD and CAD groups, EATV/height and EATV/BSA were not correlated with BMI (Figure ). When non-CAD and CAD groups were combined, EATV/BSA correlated with age in men (r = 0.320; p = 0.032), and the correlation was lost when men were segregated into non-CAD and CAD groups (Figure ). There was no correlation between EATV/BSA and age in women (Figure ). EATV/BSA correlated with the presence of diabetes mellitus in women (r = 0.431; p = 0.009) but not in men (r = 0.029; p = 0.657). EATV/BSA was not correlated with the presence of hypertension and hyperlipidemia in either men or women.
Figure 5 Linear correlation between BMI and EATV/height (upper panel) and EATV/BSA (lower panel) in men and women. Lines were plotted in non-CAD (○) and CAD subjects (●). EATV, epicardial adipose tissue volume; BSA, body surface area. Coronary (more ...)
Figure 6 Linear correlation between age and EATV/BSA in men and women. Lines were plotted in Non-CAD (○) and CAD subjects (●), respectively. EATV, epicardial adipose tissue volume; BSA, body surface area. Coronary artery disease (CAD) was defined (more ...)
Predictors for coronary atherosclerotic lesions
By univariate logistic regression analysis, presence of coronary atherosclerotic lesions, defined by >50% luminal narrowing, correlated with EATV, EATV/BSA, EATV/height, and age in men (Table ). On the other hand, the presence of coronary atherosclerotic lesions correlated with EATV and EATV/BSA in women (Table ). Body weight, BMI, systolic and diastolic blood pressure, HbA1c, triglycerides, LDL- and HDL-cholesterol, and the presence of hypertension and diabetes mellitus were not associated with the presence of coronary atherosclerotic lesions. Multivariate logistic regression analysis indicated that EATV/BSA was detected as an independent risk factor for >50% luminal narrowing only in men (Table ). BMI, age, presence of hypertension, diabetes mellitus, and hyperlipidemia were not associated with the presence of coronary atherosclerotic lesions.
Univariate analysis to estimate coronary atherosclerotic lesions in men and women
Multivariate analysis to estimate coronary atherosclerotic lesions in men and women